基于RVMD-RobustICA-ST联合相干性分析的电驱动总成噪声源识别

Noise Source Identification of Electric Drive Assembly Based on RVMDRobustICA-ST Combined with Coherence Analysis

为研究集成一体化电驱动总成的噪声源特性,提出一种RVMD-RobustICA-ST联合算法融合相干性分析的噪声源识别方法。首先,采用基于奇异值分解的占优特征值准则估计噪声子空间维数对变分模态分解参数进行指向性条件约束,并利用鲁棒性独立分量分析联合占优特征值约束的变分模态分解对信号特征进行提取。然后,利用S变换和快速傅里叶变换对各分量信号时频特性进行识别。最后,在二次残差法分析分量信号波形误差度基础上,以电驱动总成振动信号、噪声信号、时频重叠分量信号为变量建立线性系统并进行相干性分析。结果表明,稳态工况下减速二级齿副啮合振动噪声对该电驱动总成噪声贡献度最大,且时频重叠分量信号的噪声能量主要由减速一级齿副啮合振动提供。

In order to study the noise source characteristics of integrated electric drive assembly,this paper proposed a joint algorithm of RVMD-RobustICA-ST combined with coherence analysis method.Firstly,the spatial dimension of the noise was estimated by Rule of Dominant Eigenvalues(RDE)based on Singular Value Decomposition(SVD),and the variational modal decomposition was subjected to targeted conditional constraints,and then Robust Independent Component Analysis(RobustICA)combined with variational modal decomposition to extract signal features.Secondly,S Transform and Fast Fourier Transform were used to identify the time-frequency characteristics of each component signal.Finally,on the basis of analyzing the waveform error of the component signal using the quadratic residual method,a linear system was established with the vibration signal,noise signal,and time-frequency overlapping component signal of the electric drive assembly as variables for coherence analysis.The results show that the meshing vibration noise of the second gear pair of the reducer has the greatest contribution to the noise of the electric drive assembly under steady state conditions,and the noise energy of the timefrequency overlapping component signal is mainly provided by the meshing vibration of the first gear pair of the reducer.